22-10-2014, 03:57 PM
Multiresolution Analysis Using Wavelet, Ridgelet,
and Curvelet Transforms for Medical Image Segmentation
1408246947-curvelet.zip (Size: 1.84 MB / Downloads: 15)
Introduction
In the last decade, the use of 3D image processing has been increased especially for medical applications; this leads to increase the qualified radiologists’ number who navigate, view, analyse, segment, and interpret medical images. The analysis and visualization of the image stack received from the acquisition devices are difficult to evaluate due to the quantity of clinical data and the amount of noise existing in medical images due to the scanners itself. Computerized analysis and automated information systems can offer help dealing with the large amounts of data, and new image processing techniques may help to denoise those images.
Multiresolution analysis (MRA) [1–3] has been success-fully used in image processing specially with image segmen-tation, wavelet-based features has been used in various appli-cations including image compression [4], denoising [5], and classification [6]. Recently, the finite ridgelet and curvelet transforms have been introduced as a higher dimensional MRA tool [7, 8].
Image segmentation requires extracting specific features from an image by distinguishing objects from the back-ground. The process involves classifying each pixel of an
image into a set of distinct classes, where the number of classes is much smaller. Medical image segmentation aims to separate known anatomical structures from the background such cancer diagnosis, quantification of tissue volumes, radi-otherapy treatment planning, and study of anatomical struc-tures.
Segmentation can be manually performed by a human expert who simply examines an image, determines borders between regions, and classifies each region. This is perhaps the most reliable and accurate method of image segmenta-tion, because the human visual system is immensely complex and well suited to the task. But the limitation starts in volumetric images due to the quantity of clinical data.
Curvelet transform is a new extension of wavelet trans-form which aims to deal with interesting phenom-ena occurring along curved edges in 2D images [9]. It is a high-dimensional generalization of the wavelet transform designed to represent images at different scales and dif-ferent orientations (angles). It is viewed as a multiscale pyramid with frame elements indexed by location, scale, and orientation parameters with needle-shaped elements at fine scales. Curvelets have time-frequency localization properties of wavelets but also shows a very high degree of
directionality and anisotropy, and its singularities can be well approximated with very few coefficients.
This paper is focusing on a robust implementation of MRA techniques for segmenting medical volumes using features derived from the wavelet, ridgelet, and curvelet transforms of medical images obtained from a CT scanner. The rest of this paper is organised as follow: Section 2 illustrates the proposed medical image segmentation system using MRA techniques. The mathematical background and the methodology for the proposed MRA techniques have been explained in Section 3. The results and analysis of the implemented wavelet, ridgelet, and curvelet transforms for medical image segmentation are illustrated in Section 4. Finally, Section 5 includes the conclusions and future work of this research.
2. Proposed Medical Image Segmentation System
The main aim of this research is to facilitate the process of highlighting ROI in medical images, which may be encap-sulated within other objects or surrounded by noise that make the segmentation process not easy. Figure 1 illustrates the proposed medical image segmentation system using MRA. Wavelet, ridgelet, and curvelet transforms are applied on medical images with other pre- and postprocessing techniques to present segmented outputs and detected ROI in an easier and more accurate way.
[b]Methodology—Multiresolution Analysis[/b]
Image segmentation using MRA such as wavelets has been widely used in recent years and provides better accuracy in segmenting different types of images. Many recent develop-ments in MRA have taken place, while wavelets are suitable for dealing with objects with point singularities. Wavelets can only capture limited directional information due to its poor orientation selectivity. By decomposing the image into a series of high-pass and low-pass filter bands, the wavelet transform extracts directional details that capture horizontal, vertical, and diagonal activity. However, these three linear directions are limiting and might not capture enough directional information in noisy images, such as medical CT scans, which do not have strong horizontal, vertical, or diagonal directional elements. Ridgelet improves MRA segmentation; however, they capture structural information of an image based on multiple radial directions in the frequency domain. Line singularities in ridgelet transform provides better edge detection than its wavelet counterpart. One limitation to use ridgelet in image segmentation is that ridgelet is most effective in detecting linear radial structures, which are not dominant in medical images. The curvelet transform is a recent extension of ridgelet transform that overcome ridgelet weaknesses in medical image segmentation. Curvelet is proven to be particularly effective at detecting image activity along curves instead of radial directions which are the most comprising objects of medical images.
Results and Analysis
The end users of the proposed system are the radiologists and specialists who analyse medical images for cancer diagnosis. After several meetings with those people in the radiology departments in some hospitals, the main goal that they are working is to detect the accurate cancer size in medical images with the least error. This process may be affected by the noise surrounding ROI, which make the process of measuring the exact dimensions of the lesion so hard.
Different datasets have been carried out with the pro-posed system to validate it for clinical applications. The first one is NEMA IEC body phantom which consists of an elliptical water filled cavity with six spherical inserts suspended by plastic rods of inner diameters: 10, 13, 17, 22, 28, and 37 mm [25, 26]. Real clinical human images acquired by a CT scanner [24] have also been used to experiment the
proposed approaches, this data has been previously analysed by the radiologists and the provided reports explains that the patients are diagnosed by cancer. Table 1 illustrates the SNR values of extracted features from NEMA IEC DATA SET in spatial domain, different levels of decomposition of wavelet domain and at different block sizes in ridgelet domain.
It can be seen from Table 1 that small values of SNR have been obtained for all techniques; this is due to the noise from the acquisition systems itself. This noise will be a part of the medical image after the reconstruction of all slices. Relatively, better SNR values can be achieved with the second level of wavelet decomposition and as the block size (p) is getting bigger with the ridgelet transform, where the transformed image is getting more similar to the original image. This can be assigned to the major limitation of using ridgelet transformation in medical image segmentation, where ridges rarely exist in such data.
MRA transforms have been used with thresholding technique to segment the experimental data. Thresholding technique has been applied as a preprocessing step on the original images at threshold value (t = 35) to remove as much artificial spam sequel produced from the scanners. The transform then applied to effectively represent objects with edges which are the contours of the medical images followed by another thresholding at (t = 7) to remove most of the remaining noise and facilitate the measurement process.
Figure 14 illustrates the segmentation using curvelet transform. Figures 14(a) and 14© illustrate the original images from a CT scanner, and Figures 14(b) and 14(d) illustrate the segmented phantom image and real chest image, respectively, using curvelet transform. As illustrated in Figure 15, results of the proposed segmentation technique are vary in terms of smooth reconstruction of the spheres. Curvelet transform segments the input image and removes artifacts from the image to exhibit smooth and optimal segmentation of NEMA phantom. Ridgelet transform detect ROI but does not give promising segmentation results due to the lack of ridges or straight lines in the tested data set. Wavelet quadrants are varying also in their quality; relatively, the best results have been achieved with the LL-filter output.
Table 2 illustrates NEMA spheres diameters error per-centages measured using different multiresolution analysis techniques and compared to previously implemented tech-niques. ED has been used to measure the spheres diameters and calculate the error percentages for each technique, and sphere diameter error percentages have been calculated as follows: